The City of South Bend requested an assessment of the city’s current housing situation from both a rental and owner-occupied perspective. The research team analyzed social, demographic, and economic data available from the U.S. Census as well as residential housing data sources (e.g. Zillow). With this data, models were created to find commonalities within the data at a Census Block Group level to group them into clusters or “profiles”. The team then used these “profiles” to perform statistical and inferential analyses to illustrate similarities and differences among granular areas of the City of South Bend.
The results shown in greater detail in this paper found five clusters of similar Census Block Groups. These five clusters have unique and differentiating characteristics, and they span many areas of the greater South Bend area. While some are naturally grouped in close geographic proximity, it is enlightening to see Block Groups from geographically different areas of the city and outlying areas clustered together. This paper will explain the process used to extract the data, the methodology behind the models created to cluster the data, and the statistical analysis done on the clusters to create “profiles”.
The primary source of data for this analysis was the American Community Survey (ACS) conducted by the US Census Bureau. The ACS is an ongoing survey that provides vital information on an annual basis about the nation and its people. Information from the survey generates data that helps drive federal and state funding decisions. The data is made available via API from the US Census website, and for our analysis we used the 2017 data for St. Joseph County - which includes South Bend. The US Census divides geographic areas into two primary categories: a Census Tract and a Census Block Group. These are simple geographic perimeters defined by census officials. Tracts can be made up of one or more Block Groups. For St. Joseph County, there are a total of 76 Census Tracts and 227 Census Block Groups. The ACS provides a vast number of variables from which we selected approximately 104 variables for our analysis. We focused on using the Census Block Groups in our modeling as that provided more granular areas for potential clustering; however, Exploratory Data Analysis (EDA) was conducted on both Block Group and Tract data. See Appendix A for a full list of variables used.
The maps and data tables below show data related to all the block groups in St. Joseph’s county.
| Population | Two Person Families | Education - GED or Alternative | Household Income - $45,000 - $49,999 |
| Housing Units | Three Person Families | Education - Less Than 1 Year College | Household Income - $50,000 - $59,999 |
| Occupied Housing Units | Four Person Families | Education - More Than 1 Year Collage - No Degree | Household Income - $60,000 - $74,999 |
| Vacant Housing Units | Five Person Families | Education - Associates Degree | Household Income - $75,000 - $99,999 |
| Owner-Occupied Units | Six Person Families | Education - Bachelors Degree | Household Income - $100,000 - $124,999 |
| Renter-Occupied Units | 7 or More Person Families | Education - Masters Degree | Household Income - $125,000 - $149,999 |
| Units for Rent | 1 Person Non-families | Education - Professional School Degree | Household Income - $150,000 - $199,999 |
| Rented, Non-Occupied Units | 2 Person Non-families | Education - Doctorate Degree | Household Income - $200,000 or More |
| Units for Sale | 3 Person Non-families | Commute - Car, Truck or Van | Median Rent/Income |
| Sold, Non-Occupied Units | 4 Person Non-families | Commute - Drive Alone | Median Housing Cost |
| 25th Percentile Home Value | 5 Person Non-families | Commute - Carpool | Median Non-family Income |
| 50th Percentile Home Value | 6 Person Non-families | Commute - 2 Person Carpool | Median Rent - No Bedrooms |
| 75th Percentile Home Value | 7 or More Person Non-families | Commute - 3 Person Carpool | Median Rent - 1 Bedroom |
| Median Owner Cost - Mortgage | Labor Force | Commute - 4 Person Carpool | Median Rent - 2 Bedrooms |
| Median Owner Cost - No Mortgage | Civilian Labor Force | Commute - 5 or 6 Person Carpool | Median Rent - 3 Bedrooms |
| Median Owner Cost/Income - Mortgage | Civilian Labor Force - Employed | Commute - 7 or More Person Carpool | Median Rent - 4 Bedrooms |
| Median Owner Cost/Income - No Mortgage | Civilian Labor Force - Unemployed | Commute - Public Transportation | Median Rent - 5 or More Bedrooms |
| Median Family Income | Military Labor Force | Commute - Bus or Trolley | Median Rent |
| White Population | Non-Labor Force | Commute - Subway or Elevated Train | Aggregate Income Deficit |
| Black Population | Education - No Schooling | Commute - Railroad Train | Aggregate Income Deficit - Married |
| Native American Population | Education - Nursery School | Commute - Ferry Boat | Aggregate Income Deficit - Nonmarried |
| Asian Population | Education - Kindergarten | Commute - Taxi | Aggregate Income Deficit - No Wife |
| Pacific Islander Population | Education - First Grade | Commute - Motorcycle | Aggregate Income Deficit - No Husband |
| Other Population | Education - Second Grade | Commute - Bicycle | Median Household Income |
| Two or More Races Population | Education - Third Grade | Commute - Walk | Median Household Income - Owners |
| Median Age | Education - Fourth Grade | Commute - Other Means | Median Household Income - Renters |
| Median Age - Male | Education - Fifth Grade | Commute - Work from Home | Commute - Streetcar or Trolley Car |
| Median Age - Female | Education - Sixth Grade | Household Income - Less than $10,000 | Empolyed - No Vehicles Available |
| Households | Education - Seventh Grade | Household Income - $10,000 - $14,999 | Empolyed - 1 Vehicle Available |
| Family Households | Education - Eighth Grade | Household Income - $15,000 - $19,999 | Empolyed - 2 Vehicles Available |
| Family Households - Married | Education - Ninth Grade | Household Income - $20,000 - $24,999 | Empolyed - 3 Vehicles Available |
| Family Households - Other | Education - Tenth Grade | Household Income - $25,000 - $29,999 | Empolyed - 4 Vehicles Available |
| Non-family Households | Education - Eleventh Grade | Household Income - $30,000 - $34,999 | Empolyed - 5 Vehicles Available |
| Non-family Households - Alone | Education - Twelfth Grade - No Diploma | Household Income - $35,000 - $39,999 | |
| Non-family Households - Not Alone | Education - High School Diploma | Household Income - $40,000 - $44,999 |
The map below shows the boundaries of the census Block Groups, colored by their profile. Additionally, neighborhood boundaries from the South Bend Neighborhood Resources Connection and the South Bend city limits are overlayed on the map.
Below, the data table shows the average values of the Block Groups that are contained within each profile.
Below, the data table shows the total values of the Block Groups that are contained within each profile.
Next, geospatial analysis is conducted to find Block Groups that overlap with the Neighborhood Resource Council’s defined neighborhoods.
First, a map of the neighborhoods is created for reference.
Next, a table is created which lists each neighborhood, the block groups that overlap with that neighborhood, and the variables for those block groups.
Below, a table is created that shows the average values of the block groups that touch each neighborhood.
This section will dive deeper into each individual profile.
Below is a map of profile 1. A few key statistics for each block group are shown when that block group is clicked.
Relative to St. Joseph County, Profile 1 is characterized by:
Notable Block Groups: Block Group 2, Census Tract 115.05 is the best representation of Profile 1 due to a low median age (30.5 years), higher home values ($162,800 median home value), and the percentage of family households (40% of family households). Block Group 3, Census Tract 113.01 the most distinct block group of Profile 1 due to a very low median age (25.7 years), a low number of people per housing unit (.98), a very low percentage of family households (19%) and higher rental costs ($883).
Relative to St. Joseph County, Profile 2 is characterized by:
Based on these observations, Profile 2 can be described as block groups with higher home values, more people per home, and whose inhabitants own their own homes, have higher incomes, and tend to have families.
Notable Block Groups: Block Group 1, Census Tract 113.04 is the best representation of Profile 2 due to a higher median family income ($121,616), a higher percentage of owner-occupied housing units (94%), a higher percentage of family households (78.8%), and its higher median age (49.2 years old). Block Group 1, Census Tract 113.05 is the most distinct block group of Profile 2 due to a lower median home value ($199,700), lower median family income ($120,179), and a higher percentage of family households (86%).
Relative to St. Joseph County, Profile 3 is characterized by:
Based on these observations, Profile 3 can be described as block groups that are representative of St. Joseph County as a whole, albeit with a slightly older population that is more likely to own their homes and has a higher proportion of White residents.
Notable Block Groups: Block Group 1, Census Tract 12 is best representative of Profile 3 due to due to a higher median home value ($120,100), a median family income ($64,500) consistent with the rest of Profile 3, percentage of owner-occupied homes (91%), percentage of White residents (86%), and percentage of family households (64%). Block Group 3, Census Tract 10 is a distinct block group in Profile 3 due to lower home values (median value of $33,000), higher rental costs ($904), a lower percentage of owner-occupied housing units (11%), and a lower percentage of family households (14%), and a lower percentage of White people (77%). Block Group 1, Census Tract 7 is a distinct block group in Profile 3 due to lower home values (median home value of $73,600), a lower median age (34.8 years), a higher percentage of Black residents, and a higher percentage of vacant units (21%).
Relative to St. Joseph County, Profile 4 is characterized by:
Based on these observations, Profile 4 can be described as block groups with lower home values, lower home ownership costs with a mortgage, slightly lower rental costs, and much lower family incomes. The population of Profile 4 tends to be younger, contains a higher percentage of Black residents, a lower percentage of White residents, and is more likely to be unemployed.
Notable Block Groups: Block Group 1, Census Tract 29 is best representative in Profile 4 due to lower home values (median home value of $54,600), lower rental costs ($658), a lower median age (29.1), a lower percentage of occupied housing units (77.9%), a high unemployment rate (89%). However, this block group is not representative of Profile 4 due to its higher percentage of families (82%) and a very high percentage of Black people (52%).
Block Group 1, Census Tract 27 is an extreme representation of Profile 4 due to a lower median age (30.2 years), lower home values (median home value of $41,200), lower family incomes ($32,0310, a lower percentage of occupied housing units (43%), a high unemployment rate (18%). This block group also has a higher percentage of family households (72%) and a higher rental costs ($754).
Block Group 3, Census Tract 34 is an extreme representation of Profile 4 due to a lower median home value ($51,700), lower rental costs ($711), a lower percentage of occupied housing units (65%), a higher percentage of Black people (32%). This block group also skews older (39.2 years) and has a higher percentage of families (84%).
Block Group 4, Census Tract 22 is a distinct block group in Profile 4 due to a lower median home value ($42,000), a much lower median age (19.8 years), much higher rental costs ($977), a lower percentage of White people (46%), and a higher percentage of family households (84%). This block group also has an occupied-housing rate of 88% and a median family income of $42,083.
Block Group 1, Census Tract 26 is notable or having home values that are lower than the Profile 4 average (median home value of $60,200) yet has higher median rents ($834), has a very low median age (19.7 years), has a lower percentage of White residents (46%), and contains more families than Profile 4 does on average (73%).
Relative to St. Joseph County, Profile 5 is characterized by:
Based on these observations, Profile 4 can be described as block groups with higher home values and correspondingly higher home ownership costs with a mortgage, higher rental costs, and slightly more occupants per home whose inhabitants are more likely to own their homes. The population of Profile 5 tends to be older, with higher proportion of White residents, and a higher percentage of family households.
Notable Block Groups: Block Group 2, Census Tract 113.05 is representative of Profile 5 due to a higher median home value ($146,300), higher rental costs ($1,304), an older median age (44 years), more people per housing unit (2.66), a higher percentage of white people (85%), and a higher percentage of family households (75%).
Block Group 4, Census Tract 109 is an extreme representation of Profile 5 due to a higher median home value ($177,700), higher median age (46.3 years), and a higher percentage of owner-occupied housing units (91%).
This section will find block groups that overlap with the city limits and those that do not. The average values of the variables will be shown for in-city and out-city block groups, as well as the county as a whole.